Tumor microenvironment-based feed-forwardregulation of NOS2 in breast cancer progressionJulie L. Heineckea, Lisa A. Ridnoura, Robert Y. S. Chenga, Christopher H. Switzera, Michael M. Lizardob, Chand Khannab,Sharon A. Glynnc, S. Perwez Hussaind, Howard A. Younge, Stefan Ambsd, and David A. Winka,1

aRadiation Biology Branch, bTumor and Metastasis Biology Section, Pediatric Oncology Branch, and dLaboratory of Human Carcinogenesis, National CancerInstitute, National Institutes of Health, Bethesda, MD 20892; cProstate Cancer Institute, National University of Ireland, Galway, Ireland; and eLaboratory ofExperimental Immunology, National Cancer Institute, National Institutes of Health, Frederick, MD 21702

Edited* by Louis J. Ignarro, University of California, Los Angeles School of Medicine, Beverly Hills, CA, and approved March 17, 2014 (received for reviewJanuary 28, 2014)

Inflammation is widely recognized as an inducer of cancer pro-gression. The inflammation-associated enzyme, inducible nitricoxide synthase (NOS2), has emerged as a candidate oncogene inestrogen receptor (ER)-negative breast cancer, and its increasedexpression is associated with disease aggressiveness and poorsurvival. Although these observations implicate NOS2 as anattractive therapeutic target, the mechanisms of both NOS2induction in tumors and nitric oxide (NO)-driven cancer progressionare not fully understood. To enhance our mechanistic understandingof NOS2 induction in tumors and its role in tumor biology, we usedstimulants of NOS2 expression in ER− and ER+ breast cancer cells andexamined downstream NO-dependent effects. Herein, we show thatup-regulation of NOS2 occurs in response to hypoxia, serum with-drawal, IFN-γ, and exogenous NO, consistent with a feed-forwardregulation of NO production by the tumor microenvironment inbreast cancer biology. Moreover, we found that key indicators ofan aggressive cancer phenotype including increased S100 calciumbinding protein A8, IL-6, IL-8, and tissue inhibitor matrix metallopro-teinase-1 are up-regulated by these NOS2 stimulants, whereas in-hibition of NOS2 in MDA-MB-231 breast cancer cells suppressedthese markers. Moreover, NO altered cellular migration and chemo-resistance of MDA-MB-231 cells to Taxol. Most notably, MDA-MB-231 tumor xenographs and cell metastases from the fat pad to thebrain were significantly suppressed by NOS2 inhibition in nude mice.In summary, these results link elevated NOS2 to signals from thetumor microenvironment that arise with cancer progression andshow that NO production regulates chemoresistance and metastasisof breast cancer cells.

Inflammation is a major component of the tumor microenvi-ronment and a driving force in cancer initiation, promotion,

and progression (1–3). Epithelial cancers express markers ofinflammation that promote disease progression and drug re-sistance through evasion of cell death pathways and increasedtumor metastasis. Rapid cancer growth leads to tumor hypoxiaand nutrient deprivation, which promotes chronic inflammatoryfeed-forward signaling and selection of resistant tumors that areclinically challenging and sometimes untreatable.Several proinflammatory proteins such as COX2, NF-κB, IL-6,

IL-8, S100 calcium binding protein A8 (S100A8), and VEGF aremarkers of chronic inflammation in the tumor microenvironment.In addition, these proinflammatory mediators directly correlate withinducible nitric oxide synthase (NOS2), which is an emerging bio-marker of aggressive tumors that predicts poor survival in patientswith elevated tumor NOS2 expression (4–8). These and otherclinical studies warrant an improved mechanistic understanding ofintratumoral NOS2 regulation and endogenous NO production,which may be therapeutically beneficial.Toward this end, our laboratory and others have used NO

donors to study NO signaling in cancer. However, intratumoralNOS2 induction by components of the tumor microenvironment,which produces endogenous NO at levels that promote diseaseprogression and predict poor outcome, has not been examined.

Herein, we have used cell culture conditions that simulatechronic inflammation in the tumor microenvironment, whichinclude nutrient deprivation by serum withdrawal (SW), hypoxia,inflammatory cytokines, and NO donors to examine physiologicmechanisms of NOS2 induction and downstream effects of NOtarget activation in both estrogen receptor positive and negative(ER+ and ER−) breast cancer cells. In addition, we show in vivoeffects of NOS2 inhibition on tumor growth and metastases inmice. Using this approach, we provide evidence of NOS2 as a keydriver of feed-forward signaling that promotes chronic inflam-mation and cancer progression.

ResultsNOS2 Drives Tumor Growth and Metastasis. The effects of NOS2inhibiton via amminoguanidine (AG) on tumor progression wasexamined in a xenograph model of green fluorescent protein-tagged MDA-MB-231 (MDA-MB-231-GFP) breast cancer cellsimplanted in the mammary fat pad of female nude mice. Fig. 1demonstrates significantly reduced tumor growth in AG-treatedmice (Fig. 1A); after 37 d of treatment, AG suppressed tumorgrowth by 59% compared with control (44.9 ± 8.8 mm3 vs. 109 ± 17mm3, respectively). These results are supported by GFP fluores-cence imaging shown in Fig. 1B, which provides a visual index ofreduced tumor cell proliferation in AG-treated mice comparedwith control animals.The effect of NOS2 inhibition on metastatic potential of

MDA-MB-231-GFP tumors was also examined. MDA-MB-231-GFPcell brain metastasis at day 45 after fat pad injection was quan-tified by real-time PCR and fluorescence imaging of the GFPtag. Fig. 1 C and D shows a dramatic reduction in the quantifiedmRNA and fluorescent protein levels, respectively of GFP fromthe metastasized MDA-MB-231-GFP cells in the brains of

Significance

More than 92% of ER− breast cancer patients die with mod-erate to high NOS2. In this report, we show that tumor cellNOS2, through formation of a specific flux of NO, drives ER−

disease to a more aggressive phenotype. Correlation betweenNOS2-related genes from patient cohorts, in vitro, and animalexperiments show an inflammatory loop mediated by NOS2that makes ER− breast cancer an aggressive disease.

Author contributions: J.L.H., L.A.R., S.A., and D.A.W. designed research; J.L.H., L.A.R.,R.Y.S.C., and M.M.L. performed research; R.Y.S.C., M.M.L., C.K., S.P.H., and H.A.Y.contributed new reagents/analytic tools; J.L.H., L.A.R., R.Y.S.C., C.H.S., M.M.L., C.K.,S.A.G., H.A.Y., and S.A. analyzed data; and J.L.H., L.A.R., S.A.G., S.A., and D.A.W.wrote the paper.

The authors declare no conflict of interest.

*This Direct Submission article had a prearranged editor.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. E-mail: wink@mail.nih.gov.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1401799111/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1401799111 PNAS | April 29, 2014 | vol. 111 | no. 17 | 6323–6328

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

June

14,

202

0

AG-treated mice compared with control animals. Together, theseresults demonstrate that NOS2 inhibition dramatically reducestumor growth and metastasis and provides evidence that NOS2 isa key driver of breast cancer disease progression in this model.ER− patients with high NOS2 tumor levels were found to have

a gene signature that was predictive of outcome (4). Genes in-cluded in in this signature were IL-8, IL-6, S100A8, CD44, andToll-like receptor 4 (TLR4), all of which are activated by NOdonors in ER− cell lines. To investigate the effect of AG on this

gene signature, we analyzed their expression in dissected tumorsfrom this model by RT-PCR. As seen in Fig. 1E, the mRNA ex-pression of COX2, TLR4, S100A8, CD44, IL-6, and IL-8 all sig-nificantly decrease in AG-treated mice compared with control.Therefore, this model mimics the patient data and shows NOS2inhibition reduces markers that are associated with poor outcome.

Cell Migration. Earlier reports have demonstrated NO-inducedmigration of ER− breast cancer cells (4). To examine a role ofNO in MDA-MB-231 cell migration induced by SW, we used thexCelligence RTCA instrument that measures cell movementin real time. The cells were plated in serum-free RPMI 1640 in thepresence of L-Arg and allowed to adhere overnight. AG was addedwith or without varying concentrations of the NO donor 1-[N-(2-Aminoethyl)-N-(2-ammonioethyl)amino]diazen-1-ium-1,2-diolate (DETA/NO, 100 μM or 300 μM), and real-time migrationof the cells to serum containing media was monitored for 60 h.Fig. 1F shows abolished cell mobility by AG compared with con-trol cells. Titration of 100 μM DETA/NO reversed the effects ofAG and increased cell migration of AG-inhibited cells over 24 h,which returned to baseline levels and is consistent with the tem-poral release of 100 nM steady-state NO flux released by 100 μMDETA/NO under the same conditions (Fig. S1). In contrast,300 μM DETA/NO markedly increased cell migration that wassustained through 60 h. These results suggest NO flux-dependentregulation of cell migration. Interestingly, SW also inducedNOS2 expression, nitrite production (Fig. S2 A and B), and se-cretion of prometastatic cytokines such as IL-6 and IL-8, whichwas partially inhibited by AG (Fig. S3).

Drug Resistance. Chemoresistance is a recurring clinical problemin cancer treatment. Because NOS2 predicts poor cancer sur-vival, its potential contribution to chemoresistance to Taxol wasexamined. MDA-MB-231 cells were pretreated by SW (24 h)with or without AG or DETA/NO, and then exposed to Taxol(10 nM) for 18 h. The cells were then trypsinized and plated forsurvival. Compared with the untreated control, SW abated Taxolcell killing, which was augmented by 100 μM DETA/NO (Fig.1G). These results indicate NO-mediated resistance to breastcancer cell killing by Taxol under conditions of nutrient depri-vation. Collectively, these results show promotion of tumorgrowth, metastasis, and drug resistance by NOS2-derived NO.

Cytokine Stimulation of NOS2. Cytokines that induce NOS2 arepresent in the tumor microenvironment. Thus, cytokine-inducedNOS2 expression was examined in breast cancer cell lines. Fig.S4A demonstrates a twofold increase in NOS2 protein expres-sion in MCF-7 and MDA-MB-231 but not MDA-MB-468 breastcancer cells exposed for 24 h to a cytokine mix (CM). NOS2enzymatic activity was assessed by nitrite levels in the media byusing the Griess assay. Neither MCF-7 nor MDA-MB-231 nitritelevels changed after 24-h exposure of IFN-γ, IL-1β, TNF-α, orlipopolysaccharide (LPS) alone. However, nitrite productionsignificantly increased in MDA-MB-468 cells after 24 h of IFN-γor TNF-α (Fig. S4B). CM increased nitrite levels 2.5-, 2-, and2.8-fold in MCF-7, MDA-MB-231, and MDA-MB-468 cells, re-spectively, (Fig. S4C), which was inhibited by AG.The timing of NOS2 mRNA induction was examined for each

CM-stimulated cell line at 4, 24, and 48 h (Fig. S4D). NOS2mRNA induction in the least aggressive MCF-7 cells peaked at24 h and was increased 50-fold relative to control. The moreaggressive MDA-MB-231 cells exhibited peak NOS2 expressionat 4 h with a 170-fold increase. Interestingly, NOS2 expression inthe intermediately aggressive MDA-MB-468 cells was similar toMCF-7 cells with maximal induction at 24 h; however, the rel-ative increase in NOS2 levels was much greater and peaked at>4,000-fold above control levels. The human NOS2 gene con-tains cytokine responsive elements. Accordingly, CM stimulated

Fig. 1. NO effects on MDA-MB-231 cells in vivo and in vitro. (A) Tumorvolume of control and AG-treated GFP-tagged MDA-MB-231 tumor-bearingmice (n = 10 and 11 at day 37). (B) GFP fluorescence intensity at tumor site incontrol and AG-treated mice at day 45. (C and D) Representative brain GFPmRNA expression and GFP protein fluorescence intensity, respectively, fromMDA-MB-231-GFP cell brain metastasis at day 45 after fat pad injectionin control and AG-treated mice. (E) mRNA levels of COX2, TLR4, S100A8,CD44, IL-6, and IL-8 in tumors of control and AG-treated mice (day 45). (F)DETA/NO addition mediates cell migration. (G) Taxol drug resistance (at10 nM) in MDA-MB-231 cells treated with SW with or without AG andDETA/NO (±SEM, *P ≤ 0.05, ***P ≤ 0.001).

6324 | www.pnas.org/cgi/doi/10.1073/pnas.1401799111 Heinecke et al.

Dow

nloa

ded

by g

uest

on

June

14,

202

0

NOS2 promoter activity in MCF-7, MDA-MB-231, and MDA-MB-468 cells increased 2.2-, 4.2-, and 1.7-fold, respectively, asmeasured by a luciferase reporter assay (Fig. S4E). It should benoted that the message level was orders of magnitude higherthan protein expression and nitrite levels and has been docu-mented in other cell lines, indicating that there is a tight regu-lation of NOS2 that has not been completely elucidated (9).However, it has been shown in human DLD-1 cells that cytokinescan up-regulate miRNA-939 (10), which is able to block proteintranslation and may be more active in this cell line. The temporalprofile of NOS2 mRNA expression was also different in each cellline, further emphasizing the importance of understanding howand when NOS2 is up-regulated, and its impact on the aggressivephenotype of these different cancer cells.

Nutrient Deprivation by Serum Withdrawal. Nutrient deprivation isa common feature of rapidly growing tumors. Cancer cells adaptto nutrient deprivation by altering metabolism, nutrient uptake,angiogenesis, and autophagy (11, 12). SW of cells grown in cul-ture mimics nutrient deprivation, and we investigated the effectsof SW on NOS2 expression. Fig. S2B shows significant nitriteproduction in MCF-7, MDA-MB-231, and MDA-MB-468 cellsafter 24–48 h of SW. NOS2 mRNA levels increased for each cellline at each time point, with maximal increases of 5.0, 12, and 2.5in MCF-7, MDA-MB-231, and MDA-MB-468, respectively (Fig.S2C). NOS2 luciferase promoter activity increased by 1.8-, 4.0-and 2.1-fold, respectively, in the same cell lines (Fig. S2D). Inaddition, NOS2 protein levels increased after 48 h SW for eachcell line (Fig. S2A).

IFN-γ Stimulation. IFN-γ is a well-known mediator of NOS2 geneexpression. IFN-γ alone induced NOS2mRNA in the MDA-MB-231 cells (Fig. S5A). Moreover, 4-h stimulation with IFN-γ fol-lowed by SW increased nitrite accumulation at 48 h (Fig. S5B).NOS2 protein levels increased in response to 4-h IFN-γ treat-ment followed by SW compared with control (Fig. S5C). Incontrast, type 1 IFN-α or IFN-β failed to up-regulate NOS2mRNA or protein in these cells, indicating that NOS2 inductionin breast cancer cells is specific to type 2 IFN-γ (Fig. S5 B and C).

Hypoxia. Hypoxia activates various oncogenic and metabolicpathways (13) and is a hallmark of aggressive tumors (14).Hypoxia increases HIF-1α stabilization and accumulation andregulates important inflammatory genes, including NOS2, whichincreased in MCF-7 (15) and MDA-MB-231 cells (16) underhypoxic conditions. We investigated the effect of hypoxia onNOS2 protein expression by incubating breast cancer cell linesunder 1% O2 with or without SW for 24 h. As seen in Fig. S5D,protein levels increased in each cell line after 24-h hypoxiacompared with cells maintained in room air and this increaseoccurred only in the absence of serum, indicating that cellularstress in conjunction with hypoxia is required for a prolongedNOS2 up-regulation.Together, these results show that hypoxia and nutrient depri-

vation, conditions that arise in the tumor microenvironment withdisease progression and further promote disease progression andmetastatic spread, induce increased NOS2 expression and ac-tivity in human breast cancer cells.

Regulation of NOS2 by AG.Nitric oxide participates in positive andnegative feedback regulation of NOS2 through diverse mecha-nisms involving translation (17, 18), posttranslational modifica-tion, and enzymatic activity (19). The affect of AG on NOS2induction was examined by confocal microscopy of IFN-γ/SW-treated MDA-MB-231 cells. Fig. 2A demonstrates basal NOS2protein fluorescence in the SW control, which significantly in-creased in both cytoplasmic and nuclear compartments afterIFN-γ treatment. Whereas the addition of AG minimally effected

NOS2 protein expression in IFN-γ treated cells, AG treatmentdramatically reduced baseline NOS2 protein expression comparedwith SW-treated MDA-MB-231 cells (Fig. 2 A and B), whichsuggests that basal levels of NOS2-derived NO mediate feed-for-ward NOS2 protein regulation. To examine a potential role of NOin feed-forward NOS2 regulation, DETA/NO was titrated backinto AG-inhibited MDA-MB-231 cells. Titration with 300–500 μMDETA/NO increased NOS2 protein (Fig. 2B) and mRNA (Fig.2C) expression to control levels in AG-inhibited MDA-MB-231cells. Furthermore, IFN-γ stimulated cells exhibited a 50% re-duction of NOS2 mRNA after incubation with AG (Fig. 2C).Together, these results show a requirement of low NO flux(200–500 nM steady-state NO) for feed-forward NOS2 regu-lation in aggressive MDA-MB-231 breast cancer cells.

Downstream Targets of NOS. High-NOS2 ER− breast tumors ex-hibit increased expression of biomarkers, which predict poorbreast cancer survival including IL-8, IL-6, TLR4, and S100A8(4). To further examine NO regulatory mechanisms of thesebiomarkers under conditions simulating a tumor microenviron-ment, MDA-MB-231 cells were treated with IFN-γ for 4 h, fol-lowed by 48 h of SW with or without AG. Fig. 3 A and B showsinduction of S100A8 and IL-6 by INF-γ and SW that is abatedby AG. The effect of exogenous NO on AG inhibited S100A8expression was also examined. Addition of DETA/NO (Fig. 3C)increased S100A8 mRNA levels in AG-inhibited MDA-MB-231cells in a biphasic manner with maximal expression occurring at300 μM DETA/NO (∼400 nM steady-state flux). The endoge-nous receptor for S100A8 is TLR4. AG abated TLR4 basal ex-pression and that of IL-8 (Fig. 3 D and E). These results suggesta role of NO in the regulation and maintenance of these keytumor biomarkers.Tissue inhibitor matrix metalloproteinase-1 (TIMP1) is an

additional biomarker that predicts poor breast cancer patientsurvival and correlates with increased PI3k/Akt activation in

Fig. 2. Confocal microscopy of NOS2 protein expression in MDA-MB-231–treated cells. (A) SW and IFN-γ with or without AG-treated cells. (B) SW withor without AG or 300 μMDETA/NO. (C) AG decreases NOS2mRNA expressionin SW and IFN-γ–treated cells and titration of DETA/NO reestablishes NOS2mRNA expression to control levels in SW+AG treated cells (±SEM, **P ≤0.01). (Scale bars: 20 μm.)

Heinecke et al. PNAS | April 29, 2014 | vol. 111 | no. 17 | 6325

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

June

14,

202

0

patients with high NOS2 tumor expression (20). TIMP1 ex-pression and tyrosine nitration (3NT) was examined under con-ditions of IFN-γ/SW with or without AG. Fig. 3F shows IFN-γ/SW-induced TIMP1 protein that was abated by AG. Also, immu-noprecipitated TIMP1 protein exhibited increased 3NT levels.Toward this end, we have reported maximal TIMP1 nitration attwo tyrosine residues under conditions of ∼400 nM steady-stateNO flux, which maximally activated PI3k/Akt signaling throughinteraction with CD63 (20).

DiscussionElevated tumor NOS2 predicts poor survival in breast and othertypes of cancer (4, 5, 21, 22). We used murine tumor xenographsand cell culture conditions that mimic an aggressive tumor mi-croenvironment including inflammation (cytokines), nutrientdeprivation (SW), and hypoxia to examine pathways leadingto tumor NOS2 expression and downstream targets of NOS2-derived NO that predict aggressive tumor phenotypes (4). Usingthis approach coupled with pharmacological NOS2 inhibition,we provide mechanistic evidence strongly supporting a role ofNOS2-derived NO in breast cancer disease progression as de-fined by enhanced tumor biomarker expression, tumor growth,and metastatic burden, which are abated by NOS2 inhibition.The mRNA of predictive biomarkers found in high NOS2 ER−

breast tumors (COX2, TLR4, S100A8, CD44, IL-6, and IL-8)were decreased by NOS2 inhibition in our mouse model. Col-lectively, our results implicate NOS2 as a key driver of cancerprogression toward metastatic disease.In this report, components of the tumor microenvironment

promote NOS2 expression as confirmed by the assessment ofincreased NOS2 mRNA, protein, and nitrite production. Underthese conditions, we show up-regulation of NO-targeted bio-markers (IL-6, IL-8, S100A8/TLR4, and TIMP1) of cancer pro-gression (4, 20) that promote tumor cell survival, proliferation,and migration, which are suppressed by NOS2 inhibition. Wealso show intracellular mechanisms of NO-mediated feed-forward NOS2 regulation. Together, these results demonstratethat the tumor microenvironment comprises an atmospherewell suited for NOS2 expression and NOS2-derived NO (Fig. 4),

which are important for maintenance and progression of ag-gressive tumor phenotypes.The less aggressive MCF-7 breast cancer cell model generated

significant nitrite levels in response to a CM, but not by cytokinesadministered as single agents (23). Herein, we show IFN-γ–induced NOS2 and nitrite accumulation in metastatic MDA-MB-231 breast cancer cells. Despite its role in tumor surveil-lance, the diverse functions of IFN-γ in cancer are beginning toemerge because of unbalanced IFN/JAK/STAT signaling, and itis postulated to suppress tumor immune surveillance leadingto the selection of more aggressive, clinically resistant tumors(24–26). In support of these findings, distinct aggressive tumorphenotypes expressing high IFN-γ predict poor patient survival(27). Also, elevated serum IFN-γ predicted cancer recurrence andreduced disease-free survival in melanoma patients (28). BecauseINF-γ up-regulates NOS2 in aggressive breast cancer cells, weexplored potential mechanistic effects linking IFN-γ signaling withelevated NOS2 tumor expression that predicted poor breast cancerspecific survival in ER− patients (4). To accomplish this mechanisticinvestigation, we examined the promoters of 44 genes up-regulatedin high NOS2 expressing tumors (4). Among these genes, ouranalysis identified IFN regulatory factor-1 (IRF1) binding sites inthe promoters of the stem cell biomarker CD44 and the basal-likebreast cancer biomarker S100A8. IRF1 is also necessary for tran-scriptional activation of the NOS2 gene (29), and this findingimplicates an IFN signature as a key mediator of NOS2 expressionand downstream effects in these aggressive breast tumors (4).IL-8 and IL-6 are two additional biomarkers of disease pro-

gression identified in high NOS2 expressing ER− breast tumors(4) and triple negative breast cancer patients (30). Herein, weshow IFN-γ induced the mRNA expression of IL-6 that wasabated by NOS2 inhibition via AG. Circulating IL-6 is elevatedin cancer patients and known to increase STAT3 and initiatedownstream binding of STAT and NF-κB in the NOS2 pro-moter. This scenario provides an additional mechanism for feed-forward NOS2 regulation within the tumor microenvironment.IFN-γ also induced S100A8 in MDA-MB-231 cells, which wasalso abolished by AG. Interestingly, AG-inhibited S100A8 mRNAwas restored to IFN-γ–induced levels by the addition of 300 μMDETA/NO. Baseline expression of the S100A8 receptor TLR4 andIL-8 were suppressed by AG, indicating a tight regulation of theseproteins by basal levels of NOS2-derived NO in MDA-MB-231cells. Collectively, these results support a role of IFN-γ within thebreast cancer tumor microenvironment that promotes prosurvivalmechanisms in specific breast cancer disease states and implicatesthe IFN-γ/JAK/STAT pathway as a potential therapeutic target inER− breast cancer patients with high NOS2 tumor expression.Further elucidation of these and other signatures should improvepersonalized therapeutic efficacy (26).

Fig. 3. NO effects on biomarker expression in MDA-MB-231 cells treatedwith IFN-γ (4 h) and SW (48 h). (A–E ) mRNA expression of S100A8, IL-6, IL-8,and TLR4 induced by IFN-γ. (F ) TIMP1 protein expression (Upper) and ni-tration (3NT blot) of TIMP1 immuno-precipitates (Lower) in IFN-γ–stimulated cells treated with or without AG (±SEM, *P ≤ 0.05, **P ≤ 0.01,and ***P ≤ 0.001).

Fig. 4. Summary of NOS2 induction and its downstream signaling in ER−

breast cancer.

6326 | www.pnas.org/cgi/doi/10.1073/pnas.1401799111 Heinecke et al.

Dow

nloa

ded

by g

uest

on

June

14,

202

0

Nutrient deprivation (SW) and hypoxia are also commonfeatures of tumors and are important when considering theregulation of inflammatory proteins associated with poor survival(31, 32). Hypoxia increases NOS2 and COX2 mRNA in MDA-MB-231 (16) and NOS2 mRNA in MCF-7 cells (15). Our studiesshow NOS2 protein up-regulation in breast cancer cells after24-h incubation in 1% O2. Moreover, SW increased cell migra-tion of MDA-MB-231 cells at 1% O2 (33), and our results(Fig. 2F) implicate NOS2 regulation as a key driver leading totumor cell migration under these conditions. Herein, SW alonesignificantly increased NOS2 mRNA and protein in all cell lines.In tumors, nutrient deprivation results from limited vasculari-zation and forces tumor adaptation via activation of angiogenic,invasion, and survival pathways (12). NO augments tumor cellsurvival, implicating NOS2 as a mediator of tumor survival underthese conditions (4, 20, 34, 35). Importantly, the current studyshows positive feed-forward regulation of NOS2 protein andmRNA expression by exogenous NO under SW conditions, be-cause AG suppressed NOS2 levels, which then rebounded in thepresence of exogenous NO donor. We have shown that theTIMP1 protein predicted poor breast cancer survival and acti-vated Akt/BAD signaling in patients with high tumor NOS2expression (20). Also, TIMP1 nitration correlated with NO-induced PI3k/Akt/BAD prosurvival signaling in MDA-MB-231cells (20). Herein, we show IFN-γ followed by SW-induced TIMP1protein expression and nitration in MDA-MB-231 cells, which wassuppressed by AG. These results suggest that IFN-γ promotestumor survival.The regulation of NOS2 in breast tumors depends on various

conditions that are associated with the tumor microenvironmentincluding cytokines, hypoxia, nutrient deprivation, and variousmetabolic factors. This report presents evidence that NOS2 in-duction may be sustained by feed-forward regulatory loops in-volving components of the tumor microenvironment includingNO, IL-6, and COX2.

ConclusionHerein, we show that pharmacological NOS2 inhibition abatedtumor growth and metastatic burden of aggressive MDA-MB-231xenographs in mice. Moreover, low baseline NOS2 expression wasobserved in metastatic MDA-MB-231 cells grown in culture,which diminished by NOS2 inhibition, then reaccumulated in thepresence of exogenous NO. Similarly, NOS2 inhibition blockedMDA-MB-231 cell migration, which was restored to basal levelsby the addition of 100 μM DETA/NO. Importantly, NOS2inhibited cell migration was significantly enhanced above basallevels by the addition of 300 μM DETA/NO, which suggests NOflux-dependent regulation of tumor cell migration. Furthermore,NO donor pretreatment or SW suppressed Taxol cell killing ofMDA-MB-231 cells, indicating the promotion of tumor survivaland chemoresistance by NOS2-derived NO. These results suggesta requirement of feed-forward NOS2 regulation for the mainte-nance of a basal NO flux that perpetuates protumorigenic sig-naling leading to cancer progression and metastases.Various stimulants present within the tumor microenviron-

ment were found to induce NOS2 expression in vitro. Towardthis end, NOS2 was induced in MCF-7, MDA-MB-468, andMDA-MB-231; however, each cell line required differentstimulants with varied temporal profiles. Furthermore, NOS2-induced regulation of tumor biomarkers supports NOS2 as abiomarker of disease progression in breast cancer patients (4).NOS2 inhibition in MDA-MB-231 cells decreased basal IL-8 andTLR4 expression. Also, TIMP1, S100A8, and IL-6 were en-

hanced by IFN-γ in an NO-dependent manner in these cells. Thelevels of NO that activate these pathways are consistent with∼200–400 nM steady-state NO flux. These results suggest thatNOS2 can be endogenously activated and that NO donors area viable resource for understanding pathologic NO effects. Theheterogeneity of stimulation paradigms is reminiscent of that intumors and supports a role of flux-dependent NOS2-derived NOin disease progression. The elucidation of factors within the tu-mor microenvironment that regulate NOS2 and NO flux-driventumor progression could lead to more personalized therapeuticoptions for women whose breast tumors express high NOS2. Theuse of NOS2 inhibitors combined with inhibition of upstreamand downstream NO targets (i.e., IFN-γ/JAK/STAT, PI3K/AKT,NF-κB, HIF-1α) or the inhibition of microRNAs that regulatehuman NOS2, such as miRNA-939, could improve clinical out-come of breast cancer patients. Moreover, NOS2 inhibitionsuppressed growth and metastatic potential of KRAS mutantMDA-MB-231 (34, 36) cells, which suggests that NOS2 in-hibition may provide a viable option for patients with currentlyuntreatable RAS mutant tumors.

Materials and MethodsCell Culture. MCF-7, MDA-MB-231, and MDA-MB-468 human breast cancercells (American Type Culture Collection) were maintained in RPMI 1640media supplemented with 10% (vol/vol) heat inactivated FBS and 1% penicillin-streptomycin in a humidified atmosphere of 5% CO2 and room air. Cellswere plated in 60-mm dishes at densities of 2 × 106 cells in 4 mL of mediumand grown overnight (80–90% confluent). For nitrite analyses, cells wereseeded at 2 × 105 in 24-well plates and grown overnight. The medium wasreplaced with 2 mL (60-mm plates) or 0.5 mL (24-well plates) of serumand phenol-free RPMI 1640 containing penicillin/streptomycin and in-flammatory stimulants. Cytokines were added alone or as a CM of 500 U/mLIFN-γ, 20 ng/mL TNF-α, 20 ng/mL IL-1β (R&D Systems), and 20 ng/mL LPS(Sigma). IFN-α and -β (R&D Systems) were added at 500 U/mL hypoxia wasachieved by incubating the cells under 1% O2 and 5% CO2 in a Bactronhypoxic chamber (Shel Lab). NOS2 activity was inhibited by the addition of1 mM AG (Sigma), and cellular effects of NO were independently evaluatedby adding the NO donor DETA/NO (T 1/2 = 20 h at 37 °C) at various con-centrations (Larry Keefer, National Cancer Institute-Frederick). We selectedAG as the primary tool in our studies because of its previous use in clinicaltrials, its well established oral administration, and because the other commonNOS2 inhibitor, 1400W, has shown toxic effects at higher doses (37).

In Vivo Studies. Female athymic nude mice were supplied from the FrederickCancer Research and Development Center Animal Production Area(Frederick, MD). The animals were received at 8 wk of age, housed five percage, and given autoclaved food andwater ad libitum. TheNOS2 inhibitor AGwas administered in filter-sterilized drinking water at 0.5 g/L. Animals wereinjected with 750,000 GFP-tagged MDA-MB-231 human breast cancer cells inthe mammary fat pad at 9–10 wk of age and grown for 1 wk before ad-ministration of the NOS2 inhibitor AG. Tumor volume was measured bycaliper and calculated as mm3 = [width2 × length]/2, where width is thesmaller dimension and presented as mean ± SEM and GFP fluorescence andquantitation are described in SI Materials and Methods. Animal protocolswere approved and performed in accordance with principles outlined in theGuide for the Care and Use of Laboratory Animals (Institute of LaboratoryAnimal Resources, National Research Council).

Statistical Analyses. Results are presented as mean ± SEM and determinedfrom at least three independent experiments. Statistical significance wasevaluated with Student’s t tests as part of PRISM Graphpad Software andreported as *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001.

All other methods are summarized in detail in SI Materials and Methods.

ACKNOWLEDGMENTS. The work was supported by the Intramural ResearchProgram of the US National Institutes of Health, National Cancer Institute,and Center for Cancer Research.

1. Ridnour LA, et al. (2013) Molecular pathways: Toll-like receptors in the tumor microenvi-

ronment—poor prognosis or new therapeutic opportunity. Clin Cancer Res 19(6):1340–1346.2. Hussain SP, et al. (2000) Increased p53 mutation load in noncancerous colon tissue from

ulcerative colitis: A cancer-prone chronic inflammatory disease. Cancer Res 60(13):3333–3337.

3. Smith HA, Kang Y (2013) The metastasis-promoting roles of tumor-associated immune

cells. J Mol Med (Berl) 91(4):411–429.4. Glynn SA, et al. (2010) Increased NOS2 predicts poor survival in estrogen receptor-

negative breast cancer patients. J Clin Invest 120(11):3843–3854.

Heinecke et al. PNAS | April 29, 2014 | vol. 111 | no. 17 | 6327

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

June

14,

202

0

5. Grimm EA, Ellerhorst J, Tang CH, Ekmekcioglu S (2008) Constitutive intracellularproduction of iNOS and NO in human melanoma: Possible role in regulation ofgrowth and resistance to apoptosis. Nitric Oxide 19(2):133–137.

6. Lagares-Garcia JA, et al. (2001) Nitric oxide synthase as a marker in colorectal carci-noma. Am Surg 67(7):709–713.

7. Wang L, Xie K (2010) Nitric oxide and pancreatic cancer pathogenesis, prevention,and treatment. Curr Pharm Des 16(4):421–427.

8. Okayama H, et al. (2013) NOS2 enhances KRAS-induced lung carcinogenesis, in-flammation and microRNA-21 expression. Int J Cancer 132(1):9–18.

9. Chan GC, et al. (2005) Epigenetic basis for the transcriptional hyporesponsivenessof the human inducible nitric oxide synthase gene in vascular endothelial cells.J Immunol 175(6):3846–3861.

10. Guo Z, et al. (2012) miRNA-939 regulates human inducible nitric oxide synthaseposttranscriptional gene expression in human hepatocytes. Proc Natl Acad Sci USA109(15):5826–5831.

11. Garber K (2006) Energy deregulation: Licensing tumors to grow. Science 312(5777):1158–1159.

12. Izuishi K, Kato K, Ogura T, Kinoshita T, Esumi H (2000) Remarkable tolerance of tumorcells to nutrient deprivation: Possible new biochemical target for cancer therapy.Cancer Res 60(21):6201–6207.

13. Denko NC (2008) Hypoxia, HIF1 and glucose metabolism in the solid tumour. Nat RevCancer 8(9):705–713.

14. Hanahan D, Weinberg RA (2011) Hallmarks of cancer: The next generation. Cell144(5):646–674.

15. Tafani M, et al. (2010) Up-regulation of pro-inflammatory genes as adaptation tohypoxia in MCF-7 cells and in human mammary invasive carcinoma microenviron-ment. Cancer Sci 101(4):1014–1023.

16. Tafani M, et al. (2008) Induction of autophagic cell death by a novel molecule is in-creased by hypoxia. Autophagy 4(8):1042–1053.

17. Griscavage JM, Rogers NE, Sherman MP, Ignarro LJ (1993) Inducible nitric oxide syn-thase from a rat alveolar macrophage cell line is inhibited by nitric oxide. J Immunol151(11):6329–6337.

18. Park SK, Lin HL, Murphy S (1994) Nitric oxide limits transcriptional induction of nitricoxide synthase in CNS glial cells. Biochem Biophys Res Commun 201(2):762–768.

19. Albakri QA, Stuehr DJ (1996) Intracellular assembly of inducible NO synthase is limitedby nitric oxide-mediated changes in heme insertion and availability. J Biol Chem271(10):5414–5421.

20. Ridnour LA, et al. (2012) Nitric oxide synthase and breast cancer: Role of TIMP-1 inNO-mediated Akt activation. PLoS ONE 7(9):e44081.

21. Ekmekcioglu S, et al. (2000) Inducible nitric oxide synthase and nitrotyrosine in hu-man metastatic melanoma tumors correlate with poor survival. Clin Cancer Res 6(12):4768–4775.

22. Ekmekcioglu S, et al. (2006) Tumor iNOS predicts poor survival for stage III melanomapatients. Int J Cancer 119(4):861–866.

23. Loibl S, et al. (2006) Investigations on the inducible and endothelial nitric oxidesynthases in human breast cancer cell line MCF-7 - estrogen has an influence one-NOS, but not on i-NOS. Pathol Res Pract 202(1):1–7.

24. Chapat C, et al. (2013) hCAF1/CNOT7 regulates interferon signalling by targetingSTAT1. EMBO J 32(5):688–700.

25. Khodarev NN, et al. (2007) Signal transducer and activator of transcription 1 regulatesboth cytotoxic and prosurvival functions in tumor cells. Cancer Res 67(19):9214–9220.

26. Weichselbaum RR, et al. (2008) An interferon-related gene signature for DNA dam-age resistance is a predictive marker for chemotherapy and radiation for breastcancer. Proc Natl Acad Sci USA 105(47):18490–18495.

27. Zaidi MR, Merlino G (2011) The two faces of interferon-γ in cancer. Clin Cancer Res17(19):6118–6124.

28. Porter GA, et al. (2001) Significance of plasma cytokine levels in melanoma patientswith histologically negative sentinel lymph nodes. Ann Surg Oncol 8(2):116–122.

29. Martin E, Nathan C, Xie QW (1994) Role of interferon regulatory factor 1 in inductionof nitric oxide synthase. J Exp Med 180(3):977–984.

30. Hartman ZC, et al. (2013) Growth of triple-negative breast cancer cells relies uponcoordinate autocrine expression of the proinflammatory cytokines IL-6 and IL-8.Cancer Res 73(11):3470–3480.

31. Tafani M, et al. (2013) Modulators of HIF1α and NFkB in cancer treatment: Is it a ra-tional approach for controlling malignant progression? Front Pharmacol 4:13.

32. Mamlouk S, Wielockx B (2012) Hypoxia-inducible factors as key regulators of tumorinflammation. Int J Cancer 132(12):2721–2729.

33. Nagelkerke A, et al. (2013) Hypoxia stimulates migration of breast cancer cells via thePERK/ATF4/LAMP3-arm of the unfolded protein response. Breast Cancer Res 15(1):R2.

34. Switzer CH, et al. (2012) S-nitrosylation of EGFR and Src activates an oncogenic sig-naling network in human basal-like breast cancer. Mol Cancer Res 10(9):1203–1215.

35. Switzer CH, et al. (2012) Ets-1 is a transcriptional mediator of oncogenic nitric oxidesignaling in estrogen receptor-negative breast cancer. Breast Cancer Res 14(5):R125.

36. Kim RK, et al. (2013) A novel 2-pyrone derivative, BHP, impedes oncogenic KRAS-driven malignant progression in breast cancer. Cancer Lett 337(1):49–57.

37. Alderton WK, et al. (2005) GW274150 and GW273629 are potent and highly selectiveinhibitors of inducible nitric oxide synthase in vitro and in vivo. Br J Pharmacol 145(3):301–312.

6328 | www.pnas.org/cgi/doi/10.1073/pnas.1401799111 Heinecke et al.

Dow

nloa

ded

by g

uest

on

June

14,

202

0